Multi-Objective Optimization of Multi-Stage Gear Transmission Considering Scuffing Risks

Abstract Multi-objective and multi-step optimization of gear mesh is presented in this study which covers both macro and micro-geometry design variables. Two objective functions are considered including minimization of gear mesh power loss and transmission error (TE). Design constraints such as bending/contact stress limits, geometrical requirements and scuffing risk are introduced in optimization process. Scuffing is one of the most common surface failures in EV gear transmission due to high motor speed and output torque. Major causes include high contact pressure, high sliding velocity and surface temperature etc. The mainframe of its evaluation method applied in the analysis is based on AGMA standard which incorporates most tribological evaluation data needed. Some related key properties such as surface roughness and lubrication conditions are generated from more refined analysis or test results. For optimization method, a modified multi-objective particle swarm optimization (MOPSO) is applied which account for different aspects of gear performances and constraints. Results demonstrated significant reduction in both power loss and peak to peak TE at the end of iteration. Finally, the relationship among scuffing probability and other gear performance data are compared. It is indicated that optimized design with minimum power loss and TE does not have the lowest maximum surface temperature.